On 3/1/2026 2:55 PM, Van Snyder via cctalk wrote:
On Sun, 2026-03-01 at 13:12 -0600, Steve Lewis via cctalk wrote:
I still think it's a good bet that MULTICS at some point existed in
the
form of punch cards. (as also SCOPE, AOSP, and Atlas Supervisor).
IBSYS on 7090/94. PR-155 on IBM 1410/7010
I think it most likely that the master for the 1410/7010 source was on
mag tape(s), and there was some kind of card/tape oriented edit program
for changes. PR-155 proper was distributed to customers on mag tape (as
were the compilers and utilities and optional source materials), and the
OS itself was assembled/generated based on a set of macros. (I'd have
to double-check to see if there were also relocatable pre-assembled
modules involved - I think that would be likely).
Keeping such things only on cards would have been very risky, and also
very hard to deal with when implementing any kind of structured change
control, which I cannot imagine not being in place for either IBSYS or
PR-155. (Both systems also supported disk drives, though perhaps IBM
would not necessarily have used those to hold the master copy of the
source for the operating systems, compilers and utilities.)
As with IBM, I know for certain that Univac had an editor for file
"elements" - so again, it is extremely unlikely that the masters were on
card - more likely drum (FastRAND) backed up to mag tape.
Having not used the CDC operating systems, I can't speak knowledgeably
to them, but still, I can't imagine the master being on actual cards.
Way too hard to manage changes.
A story I heard about the B5500 ALGOL compiler (which evolved from the
B5000 compiler) was that its source was on mag tape, which they lost,
and had to start over - but that story might be apocryphal - they
certainly would have at least had listings.
I also saw a reference in this thread to the B8500. One of my
professors told a tale about that. UW (the real one in Wisconsin, not
the one in Washington. ;) ) had a CDC 3600 which was replaced by a
Univac 1108. There was also a bid which Burroughs won proposing the yet
to be delivered B8500. They were late and later, and provided UW with a
B5500 (which I used for my CS 302 course in XALGOL) to use in the
meantime. UW then requested a demo to ensure Burroughs was making good
progress, presumably in Detroit. They demonstrated what was ostensibly
a B8500 which performed sorta OK - but then the UW people there noticed
some odd cables. Turned out that Burroughs was actually running the
demo on a multi-processor pair of B6500 machines. The contract got
cancelled, Burroughs loaned UW a second B5500 for a while, and UW moved
from the 1108 to a Univac 1110 instead.
JRJ
